fix quotes ("" for single quote)
[MojoFacets.git] / lib / MojoFacets / Data.pm
1 package MojoFacets::Data;
2
3 use strict;
4 use warnings;
5
6 use base 'Mojolicious::Controller';
7
8 use Data::Dump qw(dump);
9 use File::Slurp;
10 use Encode;
11 use locale;
12 use File::Find;
13 use Storable;
14 use Time::HiRes qw(time);
15 use File::Path qw(mkpath);
16 use Text::Unaccent::PurePerl;
17 use Digest::MD5;
18
19 use MojoFacets::Import::File;
20 use MojoFacets::Import::HTMLTable;
21 use MojoFacets::Import::CSV;
22
23 our $loaded;
24 our $filters;
25
26 sub index {
27         my $self = shift;
28
29         my $data_dir = $self->app->home->rel_dir('data');
30         die "no data dir $data_dir" unless -d $data_dir;
31
32         my @files;
33         my $changes;
34         find( sub {
35                 my $file = $File::Find::name;
36                 if ( -f $file && $file =~ m/\.(js(on)?|txt)$/ ) {
37                         $file =~ s/$data_dir\/*//;
38                         push @files, $file;
39                 } elsif ( -f $file && $file =~ m/([^\/]+)\.changes\/(\d+\.\d+.+)/ ) {
40                         push @{ $changes->{$1} }, $2
41                 } elsif ( -d $file && $file =~ m/\.html$/ ) {
42                         $file =~ s/$data_dir\/*//;
43                         push @files, $file;
44                 } elsif ( -f $file && $file =~ m/\.csv$/i ) {
45                         $file =~ s/$data_dir\/*//;
46                         push @files, $file;
47                 } elsif ( -f $file && $file =~ m/\.storable/i ) {
48                         $file =~ s/$data_dir\/*//;
49                         push @files, $file;
50                 } else {
51                         #warn "IGNORE: $file\n";
52                 }
53         }, $data_dir);
54
55         no warnings qw(uninitialized); # mtime
56         @files = sort { $loaded->{$b}->{mtime} <=> $loaded->{$a}->{mtime} || lc $a cmp lc $b } @files,
57                         grep { defined $loaded->{$_}->{generated} } keys %$loaded;
58         my $size;
59         $size->{$_} = -s "$data_dir/$_" foreach @files;
60
61         $self->render(
62                 files => [ @files ],
63                 size => $size,
64                 loaded => $loaded,
65                 filters => $filters,
66                 dump_path => { map { $_ => $self->_dump_path($_) } @files },
67                 changes => $changes,
68         );
69 }
70
71 sub _dump_path {
72         my ( $self, $name ) = @_;
73         my $dir = $self->app->home->rel_dir('data');
74         $name =~ s/^$dir//;
75         $name =~ s/\/+/_/g;
76         return '/tmp/mojo_facets.' . $name . '.storable';
77 }
78
79 sub _save {
80         my ( $self, $path ) = @_;
81
82         my $dump_path = $self->_dump_path( $path );
83         my $first_load = ! -e $dump_path;
84         warn "save loaded to $dump_path";
85         my $info = $loaded->{$path};
86         store $info, $dump_path;
87
88         if ( $first_load ) {
89                 my $mtime = $loaded->{$path}->{mtime};
90                 utime $mtime, $mtime, $dump_path;
91                 warn "sync time to $path at $mtime\n";
92         }
93
94         warn $dump_path, ' ', -s $dump_path, " bytes\n";
95         return $dump_path;
96 }
97
98
99 sub __stats {
100
101         my $stats;
102
103         my $nr_items = $#{ $_[0] } + 1;
104
105         warn "__stats $nr_items\n";
106
107         foreach my $e ( @{ $_[0] } ) {
108                 foreach my $n ( keys %$e ) {
109                         $stats->{$n}->{count}++;
110                         my @v;
111                         if ( ref $e->{$n} eq 'ARRAY' ) {
112                                 $stats->{$n}->{array} += $#{ $e->{$n} } + 1;
113                                 @v = @{ $e->{$n} };
114                         } else {
115                                 @v = ( $e->{$n} );
116                         }
117
118                         foreach my $x ( @v ) {
119                                 $stats->{$n}->{numeric}++
120                                         if $x =~ m/^[-+]?([0-9]*\.[0-9]+|[0-9]+)$/;
121                                 $stats->{$n}->{empty}++
122                                         if length $x == 0; # faster than $x =~ m/^\s*$/;
123                         }
124
125                 }
126         }
127
128         foreach my $n ( keys %$stats ) {
129                 my $s = $stats->{$n};
130                 next unless defined $s->{array};
131                 if ( $s->{array} == $s->{count} ) {
132                         delete $s->{array};
133                         if ( $s->{count} == $nr_items ) {
134                                 warn "check $n for uniqeness\n";
135                                 my $unique;
136                                 foreach my $e ( @{ $_[0] } ) {
137                                         if ( ++$unique->{ $e->{$n}->[0] } == 2 ) {
138                                                 $unique = 0;
139                                                 last;
140                                         }
141                                 }
142                                 if ( $unique ) {
143                                         $stats->{$n}->{unique} = 1;
144                                         #warn "# $n unique ",dump( $unique );
145                                 }
146                         }
147                 }
148         }
149
150         warn "# __stats ",dump($stats);
151
152         return $stats;
153 }
154
155 sub _param_or_session {
156         $_[0]->param( $_[1] ) || $_[0]->session( $_[1] )
157 }
158
159 sub stats {
160         my $self = shift;
161         my $path = $self->_param_or_session('path');
162         warn "stats $path\n";
163         delete $loaded->{$path}->{stats};
164         $self->redirect_to( '/data/columns' );
165 }
166
167
168 sub _load_path {
169         my ( $self, $path ) = @_;
170
171         return if defined $loaded->{$path}->{generated};
172
173         my $full_path = $self->app->home->rel_file( 'data/' . $path );
174         $self->redirect_to('/data/index') unless -r $full_path;
175
176         my $dump_path = $self->_dump_path( $path );
177
178         if ( defined $loaded->{$path}->{data} ) {
179                 my $mtime = (stat($full_path))[9];
180                 return if $loaded->{$path}->{mtime} == $mtime;
181                 warn "reload $full_path, modified ", time() - $mtime, " seconds ago\n";
182         } elsif ( -e $dump_path ) {
183                 warn "dump_path $dump_path ", -s $dump_path, " bytes loading...\n";
184                 my $info = retrieve $dump_path;
185                 $loaded->{ $path } = $info;
186                 return;
187         }
188
189         my $data;
190         if ( -f $full_path ) {
191                 if ( $full_path =~ m/.csv/i ) {
192                         $data = MojoFacets::Import::CSV->new( full_path => $full_path )->data;
193                 } elsif ( $full_path =~ m/.storable/ ) {
194                         warn "open $full_path ", -s $full_path, " bytes";
195                         open(my $pipe, "<", $full_path) || die $!;
196                         while ( my $o = eval { Storable::fd_retrieve $pipe } ) {
197                                 if ( defined $o->{item} ) {
198                                         push @{ $data->{items} }, $o->{item};
199                                 } else {
200                                         warn "SKIP ",dump($o);
201                                 }
202                         }
203                         close($pipe);
204                         warn "loaded ", $#{ $data->{items} } + 1, " items from $full_path\n";
205                         $data->{generated}++;
206                 } else {
207                         $data = MojoFacets::Import::File->new( full_path => $full_path, path => $path )->data;
208                 }
209         } elsif ( -d $full_path && $full_path =~ m/.html/ ) {
210                 $data = MojoFacets::Import::HTMLTable->new( dir => $full_path )->data;
211         } else {
212                 die "can't load $full_path";
213         }
214
215         my @header;
216
217         if ( defined $data->{header} ) {
218                 if ( ref $data->{header} eq 'ARRAY' ) {
219                         @header = @{ $data->{header} };
220                 } else {
221                         warn "header not array ", dump( $data->{header} );
222                 }
223         }
224
225         my $stats = __stats( $data->{items} );
226
227         @header =
228                 sort { $stats->{$b}->{count} <=> $stats->{$a}->{count} }
229                 grep { defined $stats->{$_}->{count} } keys %$stats
230                 unless @header;
231
232         my $info = {
233                 header => [ @header ],
234                 stats  => $stats,
235                 full_path => $full_path,
236                 size => -s $full_path,
237                 mtime => (stat($full_path))[9],
238                 data => $data,
239                 defined $data->{generated} ? ( generated => 1 ) : (),
240         };
241
242         $loaded->{ $path } = $info;
243         $self->_save( $path ) unless $info->{generated};
244
245 }
246
247
248 sub load {
249         my $self = shift;
250
251         my $path = $self->param('path') || $self->redirect_to( '/data/index' );
252
253         my @paths = $self->param('paths');
254         warn "# paths ", dump @paths;
255
256         foreach my $p ( keys %$loaded ) {
257                 next if grep { /^\Q$p\E$/ } @paths;
258                 warn "remove $p from memory\n";
259                 delete $loaded->{$p};
260         }
261
262         $self->_load_path( $_ ) foreach @paths;
263
264         warn "# path $path\n";
265         $self->_load_path( $path );
266
267         $self->session( 'path' => $path );
268
269         my $redirect_to = '/data/items';
270
271         $self->session( 'header' => $loaded->{$path}->{header} );
272         if ( ! defined $loaded->{$path}->{columns} ) {
273                 my $columns_path = $self->_permanent_path( 'columns' );
274                 if ( -e $columns_path ) {
275                         my @columns = map { s/[\r\n]+$//; $_ } read_file $columns_path, binmode => ':utf8';
276                         $loaded->{$path}->{columns} = [ @columns ];
277                         warn "# columns_path $columns_path ",dump(@columns);
278                 } else {
279                         $loaded->{$path}->{columns} = $loaded->{$path}->{header}
280                 }
281
282                 $redirect_to = '/data/columns';
283         }
284         $self->session( 'columns' => $loaded->{$path}->{columns} );
285         $self->session( 'order'   => $loaded->{$path}->{columns}->[0] );
286         $self->redirect_to( $redirect_to );
287 }
288
289
290 sub _loaded {
291         my ( $self, $name ) = @_;
292         my $path = $self->session('path') || $self->param('path');
293         $self->redirect_to('/data/index') unless $path;
294
295         if ( defined $loaded->{$path}->{modified} && $loaded->{$path}->{modified} > 1 ) {
296                 my $caller = (caller(1))[3];
297                 if ( $caller =~ m/::edit/ ) {
298                         warn "rebuild stats for $path ignored caller $caller\n";
299                 } else {
300                         warn "rebuild stats for $path FORCED by modified caller $caller\n";
301 #                       $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
302                         $loaded->{$path}->{rebuild_stats} = 1;
303                         $loaded->{$path}->{modified} = 1;
304                 }
305         }
306
307         if ( defined $loaded->{$path}->{rebuild_stats} ) {
308                 warn "rebuild_stats $path";
309                 $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
310                 delete $loaded->{$path}->{rebuild_stats};
311         }
312
313         if ( ! defined $loaded->{$path}->{$name} ) {
314                 warn "$path $name isn't loaded\n";
315                 $self->_load_path( $path );
316                 if ( ! defined $loaded->{$path}->{stats} ) {
317                         warn "rebuild stats for $path\n";
318                         $loaded->{$path}->{stats} = __stats( $loaded->{$path}->{data}->{items} );
319                 }
320                 if ( ! defined $loaded->{$path}->{$name} ) {
321                         warn "MISSING $name for $path\n";
322                         $self->redirect_to('/data/index')
323                 }
324         }
325
326         $self->session( 'modified' => $loaded->{$path}->{modified} );
327
328         return $loaded->{$path}->{$name};
329 }
330
331
332 sub _checked {
333         my $self = shift;
334         my $checked;
335         $checked->{$_}++ foreach @_;
336 #       warn "# _checked ",dump($checked);
337         return $checked;
338 }
339
340 sub _permanent_path {
341         my $self = shift;
342         my $path = $self->_param_or_session('path');
343         $self->app->home->rel_dir('data') . '/' . join('.', $path, @_);
344 }
345
346 sub __unac {
347         my $n = shift;
348         $n = unac_string($n);
349         $n =~ s/\W+/_/g;
350         return $n;
351 }
352
353 sub _column_from_unac {
354         my ($self,$name) = @_;
355
356         my $stats = $self->_loaded('stats');
357         my $cols_norm = { map { __unac( $_ ) => $_ } keys %$stats };
358
359         $cols_norm->{$name} || die "can't find column $name in ", dump($cols_norm);
360 }
361
362 sub _export_path {
363         my $self = shift;
364         my $path = $self->_param_or_session('path');
365         if ( ! $path ) {
366                 warn "no path in param or session";
367                 return;
368         }
369         my $dir = $self->app->home->rel_dir('public') . "/export/$path";
370         mkpath $dir unless -e $dir;
371         my $name = __export_path_name( $path, @_ );
372         my $full = $dir . '/' . $name;
373         $full =~ s/\/+$// if -d $full; # strip trailing slash for dirs
374         return $full;
375 }
376
377 sub __export_path_name {
378         my $max_length = 80;
379
380         my $path = shift;
381         my $name = join('.', map { __unac($_) } @_ );
382         if ( length($name) > $max_length ) {
383                 $name = substr($name,0,$max_length) . Digest::MD5::md5_hex substr($name,$max_length);
384         }
385         return $name;
386 }
387
388 sub columns {
389     my $self = shift;
390
391         my $view_path = $self->_permanent_path( 'views' );
392
393         if ( $self->param('columns') ) {
394                 my @columns = $self->_param_array('columns');
395                 write_file( $self->_permanent_path( 'columns' ), { binmode => ':utf8' }, map { "$_\n" } @columns );
396                 if ( my $view = $self->param('view') ) {
397                         mkdir $view_path unless -e $view_path;
398                         write_file( "$view_path/$view", { binmode => ':utf8' }, map { "$_\n" } @columns );
399                         warn "view $view_path/$view ", -s "$view_path/$view", " bytes\n";
400                 }
401
402                 $self->redirect_to('/data/items');
403         }
404
405         if ( my $id = $self->param('id') ) {
406                 my $view_full = "$view_path/$id";
407                 if ( -f $view_full ) {
408                         my @columns = map { chomp; $_ } read_file $view_full, binmode => ':utf8';
409                         warn "view $view_full loaded ", dump @columns;
410                         $self->session( 'columns' => [ @columns ] );
411                         $self->redirect_to('/data/items');
412                 }
413         }
414
415         my $stats = $self->_loaded( 'stats' );
416
417         my @columns;
418         @columns = grep { defined $stats->{$_}->{count} } @{ $self->session('columns') } if $self->session('columns');
419
420         foreach my $c ( @{ $self->session( 'header' ) } ) {
421                 push @columns, $c unless grep { /^\Q$c\E$/ } @columns;
422         }
423
424         my @views;
425         if ( -d $view_path ) {
426                 @views = map { s{^\Q$view_path\E/*}{}; $_ } glob "$view_path/*";
427                 warn "# views ",dump @views;
428         }
429
430     $self->render(
431                 message => 'Select columns to display',
432                 stats => $stats,
433                 columns => \@columns,
434                 checked => $self->_checked( $self->_param_array('columns') ),
435                 views => \@views,
436         );
437 }
438
439 sub _param_array {
440     my ($self,$name) = @_;
441
442         my @array = $self->param($name);
443         my $path  = $self->session('path');
444
445         if ( @array ) {
446                 $self->session($name => [ @array ]);
447                 $loaded->{$path}->{$name} = [ @array ];
448         } elsif ( my $session = $self->session($name) ) {
449                 if ( ref $session eq 'ARRAY' ) {
450                         @array = @$session;
451                 } else {
452                         die "$name not array ",dump($session);
453                 }
454         }
455         #warn "# $name ",dump @array;
456         return @array;
457 }
458
459 sub _param_scalar {
460     my ($self,$name,$default) = @_;
461
462         my $scalar = $self->param($name);
463
464         if ( defined $scalar ) {
465                 $self->session($name => $scalar);
466         } else {
467                 $scalar = $self->session($name);
468         }
469
470         if ( ! defined $scalar ) {
471                 $scalar = $default;
472                 die "no default for $name" unless defined $scalar;
473                 $self->session($name => $scalar);
474         }
475
476         warn "# _param_scalar $name ",dump $scalar;
477         return $scalar;
478 }
479
480 sub filter {
481         my $self = shift;
482
483         my $name = $self->param('filter_name') || die "name?";
484         my @vals = $self->param('filter_vals');
485
486         $self->_remove_filter( $name );
487         if ( @vals ) {
488                 $self->_filter_on_data( $name, @vals );
489                 if ( my $permanent = $self->param('_permanent') ) {
490                         my $permanent_path = $self->_export_path( 'filter', $name, $permanent );
491                         write_file $permanent_path, { binmode => ':utf8' }, map { "$_\n" } @vals;
492                         warn "permanent filter $permanent_path ", -s $permanent_path;
493                 }
494         }
495
496         $self->session( 'offset' => 0 );
497
498         $self->redirect_to('/data/items');
499 }
500
501 sub _filter_on_data {
502         my ( $self, $name, @vals ) = @_;
503
504         my $path = $self->session('path');
505
506         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
507                 warn "filter $name not found in data set";
508                 return;
509         }
510
511         $filters->{$name} = [ @vals ];
512         warn "_filter_on_data $name ", $#vals + 1, " values on $path\n";
513
514         my $filter_hash;
515         $filter_hash->{$_}++ foreach @vals;
516
517         #warn "# filter_hash ",dump( $filter_hash );
518
519         my $items = $self->_loaded('data')->{items};
520
521         my $include_missing = defined $filter_hash->{_missing};
522         my $filtered_items;
523
524         foreach my $i ( 0 .. $#$items ) {
525
526                 if ( defined $items->[$i]->{$name} ) {
527                         my $row = $items->[$i]->{$name};
528                         $row = [ $row ] unless ref $row eq 'ARRAY'; # FIXME probably wrong place
529                         foreach my $v ( @$row ) {
530                                 if ( defined $filter_hash->{ $v } ) {
531                                         $filtered_items->{$i}++;
532                                 }
533                         }
534                 } elsif ( $include_missing ) {
535                         $filtered_items->{$i}++;
536                 }
537         }
538
539         #warn "# filter $name ",dump($filtered_items);
540
541         $loaded->{$path}->{filters}->{$name} = $filtered_items;
542         warn "filter $name with ", scalar keys %$filtered_items, " items created\n";
543 }
544
545
546 sub _current_filters {
547         my $self = shift;
548         my $current_filters;
549         my $stats = $self->_loaded('stats');
550
551         $current_filters->{ $_ } = $filters->{ $_ }
552                 foreach ( grep { defined $filters->{ $_ } } keys %$stats )
553         ;
554         warn "# _current_filters ",dump( keys %$current_filters );
555         return $current_filters;
556 }
557
558 sub _data_sorted_by {
559         my ( $self, $order ) = @_;
560
561         my $path = $self->session('path');
562
563         warn "_data_sorted_by $order from $path";
564
565         if ( defined $loaded->{$path}->{sorted}->{$order} ) {
566                 return $loaded->{$path}->{sorted}->{$order};
567         }
568
569         my $data = $self->_loaded( 'data' );
570         my $numeric = $self->_is_numeric($order);
571         my $missing = $numeric ? 0 : '';
572         no warnings qw(numeric);
573         my $nr = 0;
574         my @sorted = map {
575                 $_->[0]
576         } sort {
577                 if ( $numeric ) {
578                         $a->[1] <=> $b->[1]
579                 } else {
580                         $a->[1] cmp $b->[1]
581                 }
582         } map {
583                 my $v;
584                 if ( ! exists $_->{$order} ) {
585                         $v = $missing;
586                 } elsif ( ref $_->{$order} eq 'ARRAY' ) {
587                         $v = join('', @{$_->{$order}});
588                 } else {
589                         $v = $_->{$order};
590                 }
591                 [ $nr++, $v ]
592         } @{ $data->{items} }
593         ;
594
595         warn "sorted: $order numeric: $numeric items: ", $#sorted + 1, "\n";
596         #warn "# sorted ",dump( @sorted );
597
598         $loaded->{$path}->{sorted}->{$order} = [ @sorted ];
599 }
600
601
602 sub __all_filters {
603         my $order = pop @_;
604         join(',', sort(@_), 'order', $order);
605 }
606
607 our ($out, $key,$value);
608
609 sub __commit_path_code {
610         my ( $path, $i, $code, $commit_changed ) = @_;
611
612         my $items = $loaded->{$path}->{data}->{items} || die "no items for $path";
613         my $row = $items->[$i];
614         my $update;
615         eval $code;
616         foreach ( keys %$update ) {
617                 $$commit_changed->{$_}++;
618                 $loaded->{$path}->{data}->{items}->[$i]->{$_} = $update->{$_};
619         }
620         #warn "__commit_path_code $path $i ",dump( $update );
621 }
622
623 sub items {
624         my $self = shift;
625
626         if ( my $show = $self->param('id') ) {
627                 $self->param('show', $show);
628                 warn "show $show\n";
629         }
630
631         my $path = $self->_param_scalar('path');
632
633         my @columns = $self->_param_array('columns');
634         $self->redirect_to('/data/columns') unless @columns;
635         my $order   = $self->_param_scalar('order', $columns[0]);
636         my $sort    = $self->_param_scalar('sort', 'a');
637         my $offset  = $self->_param_scalar('offset', 0);
638         my $limit   = $self->_param_scalar('limit', 20);
639         $self->_param_scalar('show', 'table');
640
641         # fix offset when changing limit
642         $offset = int( $offset / $limit ) * $limit;
643
644         if ( ! grep { /^\Q$order\E$/ } @columns ) {
645                 $order = $columns[0];
646                 $self->session( order => $order );
647         }
648         my $sorted = $self->_data_sorted_by( $order );
649
650         my @filter_names;
651         if ( $filters ) {
652                 foreach my $name ( keys %$filters ) {
653                         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
654                                 warn "skip filter $name not found in $path\n";
655                                 next;
656                         }
657                         push @filter_names, $name;
658                 }
659                 warn "filter_names ",dump( @filter_names );
660                 foreach my $name ( @filter_names ) {
661                         next if ref $loaded->{$path}->{filters}->{$name} eq 'ARRAY';
662                         $self->_filter_on_data( $name, @{ $filters->{$name} } );
663                 }
664         }
665
666         my $all_filters = __all_filters( @filter_names,$order );
667
668 #       warn "# all_filters $all_filters ", dump( $loaded->{$path}->{filtered}->{$all_filters} );
669
670         if ( ! defined $loaded->{$path}->{filtered}->{$all_filters} ) {
671
672                 my $path_filters = $loaded->{$path}->{filters};
673
674                 warn "create combined filter for $all_filters from ", $#$sorted + 1, " items\n";
675
676                 my @filtered;
677                 foreach my $i ( 0 .. $#$sorted ) {
678                         my $pos = $sorted->[$i];
679
680                         if ( $#filter_names == -1 ) {
681                                 push @filtered, $pos;
682                                 next;
683                         }
684
685                         my $skip = 0;
686                         foreach ( @filter_names ) {
687                                 $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
688                         }
689                         next if $skip;
690
691                         push @filtered, $pos;
692                 }
693
694                 $loaded->{$path}->{filtered}->{$all_filters} = [ @filtered ];
695         }
696
697         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
698                 if defined $loaded->{$path}->{filtered}->{$all_filters};
699
700         warn "all_filters $all_filters produced ", $#$filtered + 1, " items\n" if $filtered;
701
702         my $data = $self->_loaded('data');
703
704         my $code = $self->_param_scalar('code','');
705         $code =~ s{\r}{}gs;
706         $code =~ s{\n+$}{\n}s;
707
708         my $commit = $self->param('commit');
709         my $test = $self->param('test');
710
711         my $commit_changed;
712
713         if ( $code && ( $test || $commit ) ) {
714                 # XXX find columns used in code snippet and show them to user
715                 my $order = 0;
716                 foreach my $column ( $code =~ m/\$row->{([^}]+)}/g ) {
717                         if ( $column =~ s/^(['"])// ) {
718                                 $column =~ s/$1$//;
719                         }
720                         next if $column =~ m/\$/; # hide columns with vars in them
721                         $commit_changed->{$column} = 0;
722                 }
723         }
724
725         my $code_path = $self->app->home->rel_dir('public') . "/code";
726         if ( $commit ) {
727
728                 warn "# commit on ", $#$filtered + 1, " items:\n$code\n";
729                 ( $key, $value, $out ) = ( 'key', 'value' );
730                 foreach ( 0 .. $#$filtered ) {
731                         my $i = $filtered->[$_];
732                         __commit_path_code( $path, $i, $code, \$commit_changed );
733                 }
734
735                 $self->_save_change({
736                         path => $path,
737                         time => $self->param('time') || time(),
738                         user => $self->param('user') || $ENV{'LOGNAME'},
739                         code => $code,
740                         commit_changed => $commit_changed,
741                 });
742
743                 if ( my $description = $self->param('code_description') ) {
744                         my $depends = $self->param('code_depends') || die "no code_depends?";
745                         my $path = "$code_path/$depends.$description.pl";
746                         if ( -e $path && ! $self->param('overwrite') ) {
747                                 warn "# code $path not saved\n";
748                         } else {
749                                 write_file(  $path, { binmode => ':utf8' }, "$code\n" );
750                                 warn "code $path ", -s $path, " bytes saved\n";
751                         }
752                 }
753
754                 # remove console
755                 $code = '';
756                 if ( $out ) {
757                         my $commit_dataset = join('.'
758                                 , $self->param('code_depends')
759                                 , $self->param('code_description')
760                                 , time()
761                         );
762                         $key ||= 'key';
763                         $value ||= 'value';
764                         warn "key $key value $value";
765                         my $items;
766                         foreach my $n ( keys %$out ) {
767                                 my $i = { $key => [ $n ] };
768                                 my $ref = ref $out->{$n};
769                                 if ( $ref eq 'HASH' ) {
770                                         $i->{$_} = [ $out->{$n}->{$_} ] foreach keys %{ $out->{$n} };
771                                 } elsif ( $ref eq 'ARRAY' ) {
772                                         $i->{$_} = $out->{$n};
773                                 } elsif ( ! $ref ) {
774                                         $i->{$value} = [ $out->{$n} ];
775                                 } else {
776                                         $i->{_error} = [ dump($out->{$n}) ];
777                                 }
778                                 push @$items, $i;
779                         };
780                         undef $out;
781                         my $stats = __stats( $items );
782                         my @columns = grep { ! m/^\Q$key\E$/ } sort keys %$stats;
783                         unshift @columns, $key;
784
785                         $loaded->{$commit_dataset} = {
786                                 header => [ @columns ],
787                                 columns => [ @columns ],
788                                 mtime => time(),
789                                 data => { items => $items },
790                                 stats => $stats,
791                                 generated => 1,
792                         };
793                         warn "# loaded out ", dump( $loaded->{$commit_dataset} );
794                         $self->session('path', $commit_dataset);
795                         $self->session('columns', [ @columns ]);
796                         $self->session('order', $key);
797                         $self->redirect_to('/data/items');
798                         return; # FIXME needed to correctly show columns
799                 }
800
801                 # this might move before $out to recalculate stats on source dataset?
802                 __path_rebuild_stats( $path );
803                 my $c = { map { $_ => 1 } @columns };
804                 my @added_columns = sort grep { ! $c->{$_} } keys %$commit_changed;
805                 warn "# added_columns ",dump( @added_columns );
806                 unshift @columns, @added_columns;
807
808                 $self->session('columns', [ @columns ]);
809                 $loaded->{$path}->{columns} = [ @columns ];
810                 warn "# new columns ",dump( @columns );
811
812                 __invalidate_path_column( $path, $_ ) foreach keys %$commit_changed;
813         }
814
815         my $sorted_items;
816         my $from_end = $sort eq 'd' ? $#$filtered : 0;
817         my $test_changed;
818         my ( $key, $value, $out ) = ( 'key', 'value' ); # XXX make local
819         foreach ( 0 .. $limit ) {
820                 my $i = $_ + $offset;
821                 last unless defined $filtered->[$i];
822                 $i = $from_end - $i if $from_end;
823                 my $id = $filtered->[$i];
824                 my $row = Storable::dclone $data->{items}->[ $id ];
825                 if ( $code && $test ) {
826                         my $update;
827                         eval $code;
828                         if ( $@ ) {
829                                 warn "ERROR evaling $@", dump($code);
830                                 $self->stash('eval_error', $@) if $@;
831                         } else {
832                                 warn "EVAL ",dump($update);
833                                 foreach ( keys %$update ) {
834                                         $test_changed->{$_}++;
835                                         $row->{$_} = $update->{$_};
836                                 }
837                         }
838                 }
839                 $row->{_row_id} ||= $id;
840                 push @$sorted_items, $row;
841         }
842
843         if ( $self->param('export') ) {
844                 my $export_path = $self->_export_path( 'items', @columns);
845                 open(my $fh, '>', $export_path) || die "ERROR: can't open $export_path: $!";
846                 print $fh "#",join("\t",@columns),"\n";
847                 foreach my $f ( 0 .. $#$filtered ) {
848                         print $fh join("\t", map {
849                                 my $i = $data->{items}->[ $filtered->[$f] ];
850                                 my $v = '\N';
851                                 if ( ! defined $i->{$_} ) {
852                                         # nop
853                                 } elsif ( ref $i->{$_} eq 'ARRAY' ) {
854                                         $v =join(',', @{ $i->{$_} });
855                                         $v = '\N' if length($v) == 0;
856                                 } elsif ( ! ref $i->{$_} ) {
857                                         $v = $i->{$_};
858                                 } else {
859                                         $v = dump $i->{$_};
860                                 }
861                                 $v;
862                         } @columns),"\n";
863                 }
864                 close($fh);
865                 warn "export $export_path ", -s $export_path, " bytes\n";
866         }
867
868         warn "# test_changed ",dump( $test_changed );
869         my $c = { map { $_ => 1 } @columns };
870         my @added_columns = sort grep { ! $c->{$_} } keys %$test_changed;
871         unshift @columns, @added_columns;
872
873         warn "# sorted_items ", $#$sorted_items + 1, " offset $offset limit $limit order $sort";
874
875         my $code_depends = $self->param('code_depends')||
876         join(',', sort grep { $test_changed->{$_} == 0 } keys %$test_changed );
877         my $code_description = $self->param('code_description') ||
878         join(',', @added_columns);
879
880         $code_depends ||= $code_description; # self-modifing
881         if ( ! $code_depends && $out ) {
882                 $code_depends = $key;
883                 $code_description = $value;
884         }
885
886         warn "# test_changed ",dump( $test_changed, $code_depends, $code_description );
887
888         $self->render(
889                 order => $order,
890                 offset => $offset,
891                 limit => $limit,
892                 sorted => $sorted_items,
893                 columns => [ @columns ],
894                 rows => $#$filtered + 1,
895                 numeric => { map { $_, $self->_is_numeric($_) } @columns },
896                 unique  => { map { $_, $self->_is_unique( $_) } @columns },
897                 filters => $self->_current_filters,
898                 code => $code,
899                 cols_changed => $commit ? $commit_changed : $test_changed,
900                 code_depends => $code_depends,
901                 code_description => $code_description,
902                 code_path => $code_path,
903                 out => $out,
904         );
905
906 }
907
908
909 sub order {
910         my $self = shift;
911         $self->session('order', $self->param('order'));
912         $self->session('sort', $self->param('sort'));
913         $self->redirect_to('/data/items');
914 }
915
916 sub _is_numeric {
917         my ( $self, $name ) = @_;
918
919         my $stats = $self->_loaded( 'stats' );
920
921         # sort facet numerically if more >50% elements are numeric
922         my $count = $stats->{$name}->{count};
923         $count   -= $stats->{$name}->{empty} if defined $stats->{$name}->{empty};
924         defined $stats->{$name}->{numeric} &&
925                 $stats->{$name}->{numeric} > $count / 2;
926 }
927
928 sub _is_unique {
929         my ( $self, $name ) = @_;
930         my $stats = $self->_loaded( 'stats' );
931         defined $stats->{$name}->{unique};
932 }
933
934 sub _remove_filter {
935         my ($self,$name) = @_;
936         warn "_remove_filter $name\n";
937
938         my $path = $self->session('path');
939
940         delete $filters->{$name};
941         delete $loaded->{$path}->{filters}->{$name};
942         warn "filters left: ", keys %{ $loaded->{$path}->{filters} };
943
944         foreach (
945                         grep { /\Q$name\E/ }
946                         keys %{ $loaded->{$path}->{filtered} }
947         ) {
948                 delete $loaded->{$path}->{filtered}->{$_};
949                 warn "remove filtered cache $_";
950         }
951 }
952
953 sub facet {
954         my $self = shift;
955
956         my $path = $self->session('path') || $self->redirect_to( '/data/index' );
957
958         if ( my $name = $self->param('remove') ) {
959                 $self->_remove_filter( $name );
960                 $self->redirect_to( '/data/items' );
961         }
962
963         my $facet;
964         my $name = $self->param('name') || die "no name";
965
966         my $all = $self->_param_scalar('all', 1);
967         my $data = $self->_loaded('data');
968
969         my $filters = $self->_current_filters;
970         my $all_filters = __all_filters( keys %$filters,$self->session('order') );
971         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
972                 if defined $loaded->{$path}->{filtered}->{$all_filters};
973
974         if ( ! $filtered || $all ) {
975                 $filtered = [ 0 .. $#{ $data->{items} } ];
976                 warn "filter all values\n";
977         } else {
978                 warn "filter using $all_filters\n";
979         }
980
981         foreach my $i ( @$filtered ) {
982                 my $item = $data->{items}->[$i];
983                 if ( ! exists $item->{$name} ) {
984                         $facet->{ _missing }++;
985                 } elsif ( ref $item->{$name} eq 'ARRAY' ) {
986                         $facet->{$_}++ foreach @{ $item->{$name} };
987                 } else {
988                         $facet->{ $item->{$name} }++;
989                 }
990         }
991
992         my $checked_values = $self->_checked( @{ $filters->{$name} } ) if defined $filters->{$name};
993
994         if ( my $code = $self->param('code') ) {
995                 my $out;
996                 foreach my $value ( keys %$facet ) {
997                         my $count = $facet->{$value};
998                         my $checked = $checked_values->{$value};
999                         eval $code;
1000                         if ( $@ ) {
1001                                 $out = $@;
1002                                 warn "ERROR: $@\n$code\n";
1003                                 last;
1004                         } elsif ( $checked != $checked_values->{$value} ) {
1005                                 warn "checked $value $count -> $checked\n";
1006                                 $checked_values->{$value} = $checked;
1007                         }
1008                 }
1009                 warn "out ",dump( $out );
1010                 $self->stash( out => $out );
1011         }
1012
1013 #       warn "# facet $name ",dump $facet;
1014
1015         my @facet_names =
1016                   $all                      ? keys %$facet
1017                 : defined $filters->{$name} ? @{ $filters->{$name} }
1018                 : keys %$facet;
1019
1020         my $numeric = $self->_is_numeric($name);
1021
1022         my $sort = $self->param('sort');
1023         # sort numeric facets with more than 5 values ascending
1024         $sort ||= $numeric && $#facet_names > 4 ? 'a' : 'c';
1025
1026         @facet_names = sort {
1027                 my $result;
1028                 if ( $sort eq 'a' ) {
1029                         $result = $numeric ? $a <=> $b : lc $a cmp lc $b;
1030                 } elsif ( $sort eq 'd' ) {
1031                         $result = $numeric ? $b <=> $a : lc $b cmp lc $a;
1032                 } elsif ( $sort eq 'c' ) {
1033                         $result = ( $facet->{$b} || -1 ) <=> ( $facet->{$a} || -1 )
1034                 } else {
1035                         warn "unknown sort: $sort";
1036                         $result = $a cmp $b;
1037                 }
1038                 $result = $a cmp $b unless defined $result; # FIXME cludge for numeric facets with invalid data
1039                 $result;
1040         } @facet_names;
1041
1042         $self->render( name => $name, facet => $facet, checked => $checked_values,
1043                 facet_names => \@facet_names, sort => $sort, numeric => $numeric,
1044         );
1045 }
1046
1047
1048 sub __invalidate_path_column {
1049         my ( $path, $name ) = @_;
1050
1051         if ( defined $loaded->{$path}->{sorted}->{$name} ) {
1052                 delete $loaded->{$path}->{sorted}->{$name};
1053                 warn "# invalidate $path sorted $name\n";
1054         }
1055
1056         foreach ( grep { m/\Q$name\E/ } keys %{ $loaded->{$path}->{filtered} } ) {
1057                 delete $loaded->{$path}->{filtered}->{$_};
1058                 warn "# invalidate $path filtered $_\n";
1059         }
1060 }
1061
1062 sub __path_modified {
1063         my ( $path, $value ) = @_;
1064         $value = 1 unless defined $value;
1065         
1066         $loaded->{$path}->{modified}  = $value;
1067
1068         warn "# __path_modified $path $value\n";
1069 }
1070
1071 sub __path_rebuild_stats { $loaded->{ $_[0] }->{rebuild_stats} = 1 };
1072
1073 sub _save_change {
1074         my ($self,$change) = @_;
1075
1076         my $change_path = $self->_permanent_path( 'changes' );
1077         mkdir $change_path unless -d $change_path;
1078         $change_path .= '/' . $change->{time};
1079         store $change, $change_path;
1080         utime $change->{time}, $change->{time}, $change_path;
1081         warn "_save_change $change_path ", dump($change);
1082 }
1083
1084 sub edit {
1085         my $self = shift;
1086         my $new_content = $self->param('new_content');
1087         $new_content  ||= $self->param('content'); # backward compatibility with old actions
1088
1089         my $i = $self->param('_row_id');
1090         die "invalid _row_id ",dump($i) unless $i =~ m/^\d+$/;
1091         my $path = $self->param('path') || die "no path";
1092         my $name = $self->param('name') || die "no name";
1093         my $status = 200; # 200 = OK, 201 = Created
1094
1095         my $data = $self->_loaded('data');
1096
1097         if ( defined $loaded->{$path}->{data}->{items}->[$i] ) {
1098                 $new_content =~ s/^\s+//s;
1099                 $new_content =~ s/\s+$//s;
1100                 my $v;
1101                 if ( $new_content =~ /\xB6/ ) { # para
1102                         $v = [ split(/\s*\xB6\s*/, $new_content) ];
1103                 } else {
1104                         $v = [ $new_content ];
1105                 }
1106
1107                 my $old = dump $loaded->{$path}->{data}->{items}->[$i]->{$name};
1108                 my $new = dump $v;
1109                 if ( $old ne $new
1110                         && ! ( $old eq 'undef' && length($new_content) == 0 ) # new value empty, previous undef
1111                 ) {
1112                         $self->_save_change({
1113                                 path => $path,
1114                                 column => $name,
1115                                 pos => $i,
1116                                 old => $loaded->{$path}->{data}->{items}->[$i]->{$name},
1117                                 new => $v,
1118                                 time => $self->param('time') || time(),
1119                                 user => $self->param('user') || $ENV{'LOGNAME'},
1120                                 unique => {
1121                                         map { $_ => $loaded->{$path}->{data}->{items}->[$i]->{$_}->[0] }
1122                                         grep { defined $loaded->{$path}->{stats}->{$_}->{unique} }
1123                                         keys %{ $loaded->{$path}->{stats} }
1124                                 },
1125                         });
1126
1127                         warn "# change $path $i $old -> $new\n";
1128                         $loaded->{$path}->{data}->{items}->[$i]->{$name} = $v;
1129
1130                         __invalidate_path_column( $path, $name );
1131
1132                         $status = 201; # created
1133                         __path_rebuild_stats( $path );
1134         
1135                         $new_content = join("\xB6",@$v);
1136
1137                 } else {
1138                         warn "# unchanged $path $i $old\n";
1139                         $status = 304;
1140                 }
1141         } else {
1142                 $new_content = "$path $i $name doesn't exist\n";
1143                 $status = 404;
1144         }
1145
1146         warn "# edit $status ", dump $new_content;
1147
1148         $self->render(
1149                 status => $status,
1150                 new_content => scalar $new_content,
1151         );
1152 }
1153
1154
1155 sub save {
1156         my $self = shift;
1157         my $path = $self->_param_or_session('path');
1158         my $dump_path = $self->_save( $path );
1159         __path_modified( $path, 0 );
1160
1161         $self->redirect_to( '/data/items' );
1162 }
1163
1164 sub export {
1165         my $self = shift;
1166
1167         my $dir = $self->app->home->rel_dir('public');
1168
1169         if ( my $import = $self->param('import') ) {
1170
1171                 if ( $import =~ m{/filter\.(.+?)\..+} ) {
1172                         my $name = $self->_column_from_unac( $1 );
1173
1174                         my @vals = map { chomp; $_ }
1175                                 read_file "$dir/export/$import", binmode => ':utf8';
1176
1177                         $self->_remove_filter( $name );
1178                         $self->_filter_on_data( $name, @vals );
1179                         $self->session( 'offset' => 0 );
1180                         $self->redirect_to('/data/items');
1181                 } else {
1182                         warn "UNKNOWN IMPORT $import";
1183                 }
1184         }
1185
1186         if ( my $remove = $self->param('remove') ) {
1187                 my $path = "$dir/export/$remove";
1188                 unlink $path if -e $path;
1189                 $path .= '.png';
1190                 unlink $path if -e $path;
1191         }
1192
1193         my $path = $self->_export_path || $self->redirect_to('/data/index');
1194
1195         my @files = grep { ! /\.png$/ } glob "$path/*";
1196         my $mtime = { map { $_ => (stat($_))[9] } @files };
1197         @files = sort { $mtime->{$b} <=> $mtime->{$a} } @files;
1198         $self->render( export => [ @files ] );
1199 }
1200
1201 sub __loaded_paths {
1202         return
1203                 grep { defined $loaded->{$_}->{data} }
1204                 keys %$loaded;
1205 }
1206
1207 1;